Helsinki University of Technology
ALOS PALSAR FOR BOREAL FOREST AND FREEZE / THAW MONITORING
Martti Hallikainen and Erkki Tomppo*
Helsinki University of Technology
*Finnish Forest Research Institute
Helsinki University of Technology
HUT Research Aircraft
y Acquired in 1994 and modified for remote sensing research
y Twin-engine turboprop aircraft Skyvan SC-7
y Sensor accommodation: Cargo bay, nose radome and port-side of fuselage (>300 kg)
y DGPS navigation system
y Min. air speed 105 km/h
Helsinki University of Technology
16-Channel Microwave Radiometer HUTRAD
y Center frequencies: 6.8, 10.65, 18.7, 23.8, 36.5, 93, and 94 GHz
y Vertical and hori-zontal polarization at each frequency
y Incidence angle 50°off nadir
y Fully polarimetric 36.5 GHz receiver (four Stokes parameters)
y Imaging 93 GHz receiver
Helsinki University of Technology
Airborne Interferometric 1.4 GHz Radiometer HUT-2D
y The first airborne interferometricmicrowave radiometer in the world
y Will be completed and tested in 2005; first flight tests with a subunit conducted in December 2003 successfully
y 36 antenna/receiver elements in U-shape of size 2 m by 2 m
y Subsystems:y Antenna/receiver elements y Local oscillator and clocky Calibrationy Correlatory Data recording
y Mass: y Instrument: 55 kgy Support frame: 30 kg
Helsinki University of Technology
HUTSCAT Scatterometer
y Helicopter-borne non-imaging FM-CW ranging scatterometer (radar), completed in 1990
y Operating frequencies 9.8 GHz (X-band) and 5.4 GHz (C-band), with VV, HH, VH, and HV polarizations at each frequency.
y Backscatter at eight channels measured quasi-simultaneously as a function of distance, 65 cm distance resolution
y HUTSCAT main applications:y Forest tree height
measurementy Forest stem volume and
biomass estimation
Helsinki University of Technology
Use of HUTSCAT Data for Forest Research
y A radar backscatter profile of boreal forest over 850 m distance (C-band, VV-pol.)yWhite: strongest
signaly Blue: weakest signaly Black: no backscatter
y Backscattering coefficienty Blue: forest
contribution (total -ground)
y Black: total value
Helsinki University of Technology
Samples of HUTSCAT Data for Forest and Land Use Research
Helsinki University of Technology
HUT Semiempirical Forest Backscattering Modely Starting point: Relationship of sigma0 with biomass, canopy
moisture and soil moisture is complex => use training areas
y Radiative transfer model, used as follows:
y 1) For the training area, the soil and vegetation moistureparameters are estimated from the backscattering coefficients and stem volume (must be known for training areas) with the semi-empirical backscattering model.
y 2) For the whole area of interest, the stem volume is estimated from the backscattering coefficients using the semi-empirical backscattering model with the moisture parameters.
y 3) If several SAR images are used, the stem volume estimates are combined with a multiple linear regression. The regression equation is defined using the stem volume estimates for the training area.
y Current Modeling Activities: Polarimetric scattering model under development (tree made up of cylinders; tree growth model)
Helsinki University of Technology
Land Use by SAR Interferometry
y Land use map of Helsinki Metropolitan Area
y 28 images from ERS-1 and ERS-2 SAR over a period of one year
y RGB imagey R = average intensityy G = average coherence of 14
ERS Tandem image pairs for discriminating forested areas
y B = average of two long-term coherences (35 and 245 days) for discriminating urban areas
y Compares favorably withground truth
Helsinki University of Technology
Land-Cover and Stem Volume Map for Tuusula Region (ERS SAR)
y Except for the used training stand information the map is completely radar-based
y Continuous stem-volume estimates are here discretized for display.
480-540
380-480
310-380
250-310
220-250
180-220
150-180
130-150
110-130
90-110
70-90
50-70
40-50
20-40
0-20
Stem
Volume
m /ha)( 3
Agri./Open Urban Water
Helsinki University of Technology
Source: EEA
Helsinki University of Technology
Land Cover and Land Use in FinlandCorine 2000
Source: Finnish Environment Institute
Helsinki University of Technology
Mean Value of Growing Stock (m3/ha) 1989-1994
y Source: Finnish Forest Research Institute
Helsinki University of Technology
Proportions of pine, spruce and birch on forest land 1989-1994
Source: Finnish Forest Research Institute
Helsinki University of Technology
Problem Statement for Boreal Forest, v1
yEstimate the areas of six land use categories (Forest, Cropland, Grassland, Wetlands, Settlements and Other land) and changes between categories as defined in the Good Practice Guidance for Land Use, Land-Use Change and Forestry (GPG-LULUCF) of the Intergovernmental Panel on Climate Change (IPCC).
y In addition to UNFCCC LULUCF reporting, the methods could be applied in identifying Deforestation, Afforestation and Reforestation defined in Kyoto Protocol.
Helsinki University of Technology
Relevance of ALOS and Problem Statement for Boreal Forest, v2
yThe low frequency and cross-pol of ALOS PALSAR provide sensitivity to biomass but probably less so to differences between the other five categories. Hence, more practical tasks for using ALOS PALSAR are:
yTest its feasibility to estimate changes between Forest and the other five land use categories
yTest its feasibility to estimate above ground living biomass
Helsinki University of Technology
Forest Definition
yForests are lands of more than 0.5 hectares, with trees able to reach a minimum height of 5 meters maturity in situ and with a canopy cover of more than 10 percent, which are not primarily under agricultural or urban land use.
Helsinki University of Technology
Prototype Area and Ground Truth
yPrototype area: Finland
y Ground truth:
yNational Forest Inventory data
yCORINE 2000 land use information
Helsinki University of Technology
Planned Output Products and Product Developers
yOutput products:
yFeasibility of ALOS PALSAR to produce maps of annual change of forest land vs. other five land use categories
yFeasibility of ALOS PALSAR to produce above ground living biomass
y Product developers:
yHelsinki University of Technology
yFinnish Forest Research Institute
Helsinki University of Technology
Source:S.R. Eyre (1968)
Helsinki University of Technology
Accuracy of Finnish DEM
yGreen: Mean error 1.76 my Pink: 1.76 – 5 myLight brown: 5-10 myDark brown <10 myBlank: no estimate
Helsinki University of Technology
Use of HUT Semiempirical Forest Backscattering Model for Freeze / Thaw Monitoring
y Instead of determining biomass / stem volume using the backscatter model, determine soil moisture
yThen, biomass / stem volume has to be known with reasonable accuracy
yOur experimental results indicate that if derived soil moisture is smaller than 9 % by volume, soil is frozen (C-band SAR data)
yDiectric constant does not saturate at 0 deg Celsius but, rather, decreases with decreasing temperature
Helsinki University of Technology
Other HUT Activities for Soil State Monitoring
yAirborne measurements with our multi-frequency HUTRAD radiometer starting in spring 2005 (C/X-band)
yGround-based measurements with L-band radiometer starting in summer 2005; later airborne measurements
ySoil state:yFrozen / thawedyEffect of land-cover categories to the observed results
y=> Determine possibilities of L-band for freeze / thaw (few results in literature; Wegmuller: good separability)
yEarlier: dielectric measurements of frozen soil by HUTyModeling of emission/backscatter based on literature
Helsinki University of Technology
Every 10 days In-Situ Observations at 38 Finnish Stations:Ground Frost and Snow Depth for Forested / Non-forested Areas
Source: Finnish Environment Institute
Helsinki University of Technology
Northern Finland:
Maximum mean
ground frost 160
cm
Source: Finnish Environment Institute
Helsinki University of Technology
Network of PhenologyObservation Stations
Source: Finnish Environment Institute
Helsinki University of Technology Observed Leafing in Spring in 2002-204
Source: Finnish Environment Institute
Helsinki University of Technology Yellowing of leaves in Fall in 2002-204
Source: Finnish Environ-mental Institute
Helsinki University of Technology
Goal Statement for Freeze / Thaw
yDevelop annual maps of landscape freeze/thaw state within the prototype areas
yProvide estimates of thaw onset and freeze-up within the temporal fidelity allowed by ALOS PALSAR coverage
ySupport development with satellite scatterometer(QuikScat for snow melt onset / end) and possibly mw. radiometer (SMOS: Launch 2007) data
Helsinki University of Technology
Prototype Area and Ground Truth
yPrototype area: Finland
y Ground truth:
yWeather data from the Finnish Met Office
yGround-based ground frost and snow data from Finnish Environment Institute’s observation stations
y Supporting data
ySatellite scatterometer data and microwave radiometer data to fill in the gaps in ALOS PALSAR data
Helsinki University of Technology
Planned Output Products and Product Developers
yOutput products:
yAnnual maps of landscape freeze / thaw status
y Product developers:
yHelsinki University of Technology
yFinnish Forest Research Institute
yFinnish Environment Institute
Helsinki University of Technology
Permafrost in Russia
Top Related